论文部分内容阅读
低NO_x排放是电站锅炉燃烧优化的主要目标之一,影响燃煤锅炉NO_x排放因素众多且复杂,对锅炉燃烧过程NO_x浓度进行准确预测是低NO_x燃烧优化的基础。机组全工况运行时表现出强时变性,静态预测模型难以保证预测精度,考虑到观测样本的时效性,模拟记忆模式对观测数据进行重采样,进而基于支持向量回归算法构建NO_x排放预测模型,构造一种基于记忆模式的支持向量回归算法。以某机组热态试验数据为基础,对算法进行了仿真分析,结果表明,该算法在保证回归建模精度的同时,在训练速度、稳定性以及泛化性能等方面较传统支持向量回归算法更有优势。
Low NO_x emission is one of the main goals of power plant boiler combustion optimization. Influencing the NO_x emission factors of coal-fired boiler is numerous and complex. Accurately predicting NO_x concentration in boiler combustion process is the basis of low NO_x combustion optimization. When the whole working condition of the unit is running, it shows strong time-variability. The static prediction model can not guarantee the prediction accuracy. Considering the timeliness of the observed samples, the simulated memory model re-samples the observed data, and then builds the NO_x emission prediction model based on the support vector regression Construct a support vector regression algorithm based on memory model. Based on the thermal test data of a certain unit, the algorithm is simulated and analyzed. The results show that the proposed algorithm is more efficient than the traditional support vector regression algorithm in terms of training speed, stability and generalization performance while ensuring the accuracy of regression modeling There are advantages.